PDA

View Full Version : Spider math...



shocker455
07-27-2015, 11:06 PM
Iv been trying to set up a excel spread sheet to calculate the odds of getting a spider to spawn if X cards are removed from my opponents deck. Hypergeometric Distribution doesn't quite cut it, because it doesn't account for the fact that every time it does it hit, it needs to increas the sample size.

Any one know what formula would work for this?

magic_gazz
07-27-2015, 11:14 PM
I wouldn't bother if I was you, the odds of hitting your spiders and ACTUALLY hitting them seem to be far apart most of the time.

Oroniss
07-27-2015, 11:41 PM
In my experience it really depends on whether you are relying on the spiders to win :).

Vorpal
07-28-2015, 10:29 AM
From my empirical evidence, the # is 0/X

One day I will win with a spider deck.

Saeijou
07-28-2015, 10:30 AM
well... just continue to calculate. if you draw a spider, you have the next chance again... with one card and one spider less...
you just have to continue the hypergeometric distribution

Svenn
07-28-2015, 11:05 AM
From my empirical evidence, the # is 0/X

One day I will win with a spider deck.

The key to spider decks is control. You need to be heavy control so you last long enough to be getting tons of spiders. Spider decks win at the long game, so anything you can do to stretch it out until the spiders just start pouring in. Things that interrupt/destroy AND add spiders are great (parriphagy/suffocate/incubation webs). Phenteo also helps. ;)

MatWith1T
07-28-2015, 11:18 AM
The most successful spider decks I've made/faced use spiders as an alternate wincon - spiders are the finishing blow on a stalled board, or free chump blockers. If a cards primary purpose is to make spiders, it's a trap...

shocker455
07-28-2015, 11:45 AM
well... just continue to calculate. if you draw a spider, you have the next chance again... with one card and one spider less...
you just have to continue the hypergeometric distribution
Easier said then done...

For example say the opponent has a deck of 28 cards, with 6 spiders eggs. If i were to cast Lunacy(mill 5, and then opponent draws a card on his turn to hit 6 cards id have the following odds.

amount - chance to get exactly - chance to get at least
0 -19.8%
1 - 41.94% - 80.2%
2 - 29.12% - 38.26%
3 - 8.18% - 9.13%
4 - .92% - .96%
5 - .04%

That doesn't take in the chance to hit another spider because you hit one. All the percents (besides 0) should be shifted up higher, but i have no idea how to calculate it. At least not in excel with out making a huge amount of columns feeding into each other. Figured with the large amount of computer/math people tcg tend to attract someone might know of a nice formula

Zophie
07-28-2015, 12:07 PM
http://i.imgur.com/ekoQHsr.gif

Saeijou
07-28-2015, 12:17 PM
Easier said then done...

For example say the opponent has a deck of 28 cards, with 6 spiders eggs. If i were to cast Lunacy(mill 5, and then opponent draws a card on his turn to hit 6 cards id have the following odds.

amount - chance to get exactly - chance to get at least
0 -19.8%
1 - 41.94% - 80.2%
2 - 29.12% - 38.26%
3 - 8.18% - 9.13%
4 - .92% - .96%
5 - .04%

That doesn't take in the chance to hit another spider because you hit one. All the percents (besides 0) should be shifted up higher, but i have no idea how to calculate it. At least not in excel with out making a huge amount of columns feeding into each other. Figured with the large amount of computer/math people tcg tend to attract someone might know of a nice formula

in that case you can use this site:
http://stattrek.com/online-calculator/hypergeometric.aspx

an explanation for what the different things mean can be found here:
http://fiveshards.com/from-the-deck-mulling-mulligans/

Saeijou
07-28-2015, 12:21 PM
nvm... i guess you are already using the site ;)

if you calculate the odds of getting more spiders, you just add one to number of successes and one to sample size!

and of course the number get's smaller afterwards, since it is more unlikely to find more spiders in 6 cards

EDIT:

so if you draw 5 spiders, your opponent will draw 10 cards. the odds are 0.012%

3733

Jormungandr
07-28-2015, 01:27 PM
Iv been trying to set up a excel spread sheet to calculate the odds of getting a spider to spawn if X cards are removed from my opponents deck. Hypergeometric Distribution doesn't quite cut it, because it doesn't account for the fact that every time it does it hit, it needs to increas the sample size.

Any one know what formula would work for this?

Hypergeometric Distribution should work just fine for this IF you're only interested in the odds of getting at least one spider to spawn. If you're interested in the expected number of spiders, etc, then it gets more complex, but essentially because the fail case involves drawing X cards, and only X cards, and the success case involves at least one of those cards being a spider, you can do a hypergeometric distribution with a sample size of X, and the number of successes in the sample of 1. Then look at the >=1 line for the chance of at least one spider.

If you're looking to see anything more detailed, the math gets trickier, as you'll need to account for additional draws, like you mention.

Biz
07-28-2015, 04:58 PM
this might be accurate

docs.google.com/spreadsheets/d/12G09sU_lOA-WZ0wh4lmJQ5UmXO6s20JBxaluDC4d0Sg